Emerging trends in software development has been changed due to the huge amount of data, growth of internet, mobile, dynamic and smart applications. These applications consist of small, intelligent, flexible and distributed components known as agents. This research proposes agent-based autonomous controller (ABAC) architecture for managing road traffic. It uses time series of historical traffic intensity to estimate the appropriate time allocation for each signal at a given intersection. Our approach takes care of the exceptional appearance of rescue vehicles (e.g., ambulance) in order to ensure a smooth flow of the traffic. The ABAC architecture counts on several AI techniques germane to assessing the intensity of the traffic using image recognition algorithms. It also counts on environment sensors (sound sensors) in order to detect the advent of emergency vehicles. The ABAC traffic management architecture shows a high degree of adaptability leading to the least need for human intervention.